BACKGROUND/AIM: Traumatic dental injuries (TDIs) significantly affect quality of life and healthcare costs. Large administrative datasets using ICD codes offer opportunities for epidemiological research; however, the accuracy of International Classification of Diseases, Tenth Revision (ICD-10) codes for TDIs remains unclear. This study aimed to evaluate the validity of ICD-10 codes in identifying TDIs and their ability to specify injury type in a pediatric emergency department (ED) setting. MATERIALS AND METHODS: A cross-sectional diagnostic accuracy study was conducted at a large urban-based children's hospital. Children aged 0-18 years presenting to the ED between January 2021 and August 2022 were included. Patients with documented dental injuries (positive controls) or facial injuries (negative controls) were identified using structured dental trauma flowsheets and computed tomography (CT) scan records, respectively. Manual chart reviews served as the reference standard. Accuracy, sensitivity, specificity, and predictive values of ICD-10 codes for TDI were calculated. Secondary analysis assessed whether ICD-10 codes matched the specific type of TDI (fracture vs. periodontal injury). RESULTS: A total of 405 patients (mean age 10.5 ± 4.7 years) were analyzed. Overall accuracy of ICD-10 codes for TDIs was 78.5%, with sensitivity of 60.1% and specificity of 98.4%. Positive predictive value was 97.6%, and negative predictive value was 69.4%. Among true positives (n = 127), 96% of ICD-10 codes correctly reflected the specific injury type (fracture vs. periodontal injury). The most frequent correct codes were S03.2XXA (dislocation of tooth, initial encounter) and S02.5XXA (fracture of tooth, initial encounter). The most common incorrect code among false negatives was S09.93XA (unspecified injury of face). CONCLUSIONS: ICD-10 codes demonstrate high specificity but moderate sensitivity for TDI in pediatric ED records. Undercoding remains a concern and may lead to underestimating TDI prevalence and incidence. Improved education and coding practices are essential to enhance data reliability for research and healthcare planning.
Azadani et al. (Wed,) studied this question.